1. François FagesShonan village 14/11/11 Formal Cell Biology in Biocham François Fages Constraint Programming Group http://contraintes.inria.fr/ INRIA Paris-Rocquencourt France http://contraintes.inria.fr/ To tackle the complexity of biological systems, investigate: Programming Theory Concepts Formal Methods of Circuit and Program Verification Constraint Modeling and Automated Reasoning Tools Implementation in the Biochemical Abstract Machine BIOCHAM v3.3

2. François FagesShonan village 14/11/11 Systems Biology Challenge Aims at system-level understanding of multiscale biological processes in terms of their elementary interactions at the molecular level. Mitosis movie [Lodish et al. 03]  Systems Biology Markup Language (SBML): model exchange format  Model repositories: e.g. biomodels.net 1000 models of cell processes  Modeling environments (Cell designer, Cytoscape, Copasi, Biocham,…) The limits on what is observable and uncertainty of data make reasoning with partial information structures necessary  symbolic, logic, constraints

3. François FagesShonan village 14/11/11 Models in Systems Biology Models are built in Systems Biology with two contradictory perspectives : 1) Models for representing knowledge : the more detailed the better 2) Models for making predictions : the more abstract the better [Kohn 1999] [Tyson 1991]  Organize models and formalisms in hierarchies of abstractions

4. François FagesShonan village 14/11/11 Formal Reaction Rules Complexation: A + B => A-B. Decomplexation A-B => A + B. cdk1+cycB => cdk1–cycB Phosphorylation: A =[K]=> A~{p}. Dephosphorylation A~{p} =[P]=> A. Cdk1-CycB =[Myt1]=> Cdk1~{thr161}-CycB Cdk1~{thr14,tyr15}-CycB =[Cdc25~{Nterm}]=> Cdk1-CycB Synthesis: _ =[G]=> A. Degradation: A =[C]=> _. _ =[#E2-E2f13-Dp12]=> RNAcycA Transport: A::L1 => A::L2. Cdk1~{p}-CycB::cytoplasm => Cdk1~{p}-CycB::nucleus

5. François FagesShonan village 14/11/11 Semantics of Reaction Rules A + B => C Discrete Semantics: numbers of molecules Multiset rewriting, Petri net A, B  C++ A- - B- - CHAM [Berry Boudol 90] [Banatre Le Metayer 86] Boolean Semantics: presence-absence of molecules Asynchronous Transition System A  B  C  A/  A  B/  B k*[A]*[B] for A + B => C Differential Semantics: concentrations Ordinary Differential Equations Hybrid Automata (ODE+events) Stochastic Semantics: numbers of molecules Continuous time Markov chain

6. François FagesShonan village 14/11/11 Hierarchy of Semantics Stochastic traces Discrete traces abstraction concretization Boolean traces Theory of abstract Interpretation Abstractions as Galois connections [Cousot Cousot POPL’77] Thm. Galois connections between the syntactical, stochastic, discrete and Boolean semantics [Fages Soliman CMSB’06,TCS’08] Reaction rules ODE traces

7. François FagesShonan village 14/11/11 Hierarchy of Semantics Stochastic traces ODE traces Discrete traces abstraction concretization Boolean traces Thm. Under appropriate conditions the ODE semantics approximates the mean stochastic behavior [Gillespie 71] Reaction rules Theory of abstract Interpretation Abstractions as Galois connections [Cousot Cousot POPL’77] Thm. Galois connections between the syntactical, stochastic, discrete and Boolean semantics [Fages Soliman CMSB’06,TCS’08]

8. François FagesShonan village 14/11/11 Influence Graph Abstraction ODE model abstraction concretization Reaction rule model Jacobian matrix Influence graph Thm. Positive (resp. negative) circuits in the influence graph are a necessary condition for multistationarity (resp. oscillations) [Thomas 81] [Snoussi 93] [Soulé 03] [Remy Ruet Thieffry 05] [Richard 07]

9. François FagesShonan village 14/11/11 Influence Graph Abstraction ODE model abstraction concretization Reaction rule model Thm. Under increasing kinetics and in absence of double positive negative influence pairs, both influence graphs are identical [Fages Soliman FMSB 08] = Stoichiometric Influence graph Jacobian matrix Influence graph Thm. Positive (resp. negative) circuits in the influence graph are a necessary condition for multistationarity (resp. oscillations) [Thomas 81] [Snoussi 93] [Soulé 03] [Remy Ruet Thieffry 05] [Richard 07]

10. François FagesShonan village 14/11/11 Reaction Graph Reductions 011_levc Computed hierarchy of MAPK models in http://www.biomodels.nethttp://www.biomodels.net [Gay Fages Soliman, A graphical method to reduce and relate models in systems biology 2010 bi] Reaction graph reductions as subgraph epimorphisms operations: delete/merge of species/reactions vertices NP-complete problem, Constraint Logic Program (GNU Prolog) refinement simplification

11. François FagesShonan village 14/11/11 Cell Cycle Control

12. François FagesShonan village 14/11/11 Mammalian Cell Cycle Control Map [Kohn 99]

13. François FagesShonan village 14/11/11 Kohn’s map detail for Cdk2 Complexations with CycA and CycE cdk2~$P + cycA-$C => cdk2~$P-cycA-$C where $C in {_,cks1}. cdk2~$P + cycE~$Q-$C => cdk2~$P-cycE~$Q-$C where $C in {_,cks1}. p57 + cdk2~$P-cycA-$C => p57-cdk2~$P-cycA-$C where $C in {_, cks1}. cycE-$C =[cdk2~{p2}-cycE-$S]=> cycE~{T380}-$C where $S in {_, cks1} and $C in {_, cdk2~?, cdk2~?-cks1} 147 rule patterns  2733 expanded rules [Chabrier Chiaverini Danos Fages Schachter 04] How to query the possible dynamical properties of such a system ? reachability, checkpoints, steady states, oscillations ?

14. François FagesShonan village 14/11/11 Temporal Logic Queries Temporal logics introduced for program verification by [Pnueli 77] Computation Tree Logic CTL [Emerson Clarke 80] Non-det. Time E exists A always X next time EX(  )AX(  ) F finally EF(  )  AG(   ) AF(  ) liveness G globally EG(  )  AF(   ) AG(  ) safety U until E (  1 U  2)A (  1 U  2)

15. François FagesShonan village 14/11/11 Kohn’s Map Model-Checking 147-2733 rules, 165 proteins and genes, 500 variables, 2 500 states. Biocham NuSMV model-checker time in seconds [Chabrier Fages CMSB 03] Initial state G2Query:Time in sec. compiling29 Reachability G1EF CycE2 Reachability G1EF CycD1.9 Reachability G1EF PCNA-CycD1.7 Checkpoint for mitosis complex  EF (  Cdc25~{Nterm} U Cdk1~{Thr161}-CycB) 2.2 Oscillations CycA EG ( (EF  CycA) & (EF CycA)) 31.8 Oscillations CycB EG ( (EF  CycB) & (EF CycB)) false ! 6

16. François FagesShonan village 14/11/11 Constraint Linear Time Logic LTL(R)

17. François FagesShonan village 14/11/11 Quantitative Model of Cell Cycle Control [Tyson 91] k1 for _ => Cyclin. k3*[Cyclin]*[Cdc2~{p1}] for Cyclin + Cdc2~{p1} => Cdc2~{p1}-Cyclin~{p1}. k6*[Cdc2-Cyclin~{p1}] for Cdc2-Cyclin~{p1} => Cdc2 + Cyclin~{p1}. k7*[Cyclin~{p1}] for Cyclin~{p1} => _. k8*[Cdc2] for Cdc2 => Cdc2~{p1}. k9*[Cdc2~{p1}] for Cdc2~{p1} => Cdc2. k4p*[Cdc2~{p1}-Cyclin~{p1}] for Cdc2~{p1}-Cyclin~{p1} => Cdc2-Cyclin~{p1}. k4*[Cdc2-Cyclin~{p1}]^2*[Cdc2~{p1}-Cyclin~{p1}] for Cdc2~{p1}-Cyclin~{p1} =[Cdc2-Cyclin~{p1}]=> Cdc2-Cyclin~{p1}.

18. François FagesShonan village 14/11/11 biocham: learn_parameter([k3,k4],[(0,200),(0,200)],20, oscil(Cdc2-Cyclin~{p1},3),150). Parameter Search from LTL(R) Properties

19. François FagesShonan village 14/11/11 Parameter Search from LTL(R) Properties biocham: learn_parameter([k3,k4],[(0,200),(0,200)],20, oscil(Cdc2-Cyclin~{p1},3),150). First values found : parameter(k3,10). parameter(k4,70).

20. François FagesShonan village 14/11/11 Parameter Search from LTL(R) Properties biocham: learn_parameter([k3,k4],[(0,200),(0,200)],20, oscil(Cdc2-Cyclin~{p1},3) & F([Cdc2-Cyclin~{p1}]>0.15), 150). First values found : parameter(k3,10). parameter(k4,120).

21. François FagesShonan village 14/11/11 biocham: learn_parameter([k3,k4],[(0,200),(0,200)],20, period(Cdc2-Cyclin~{p1},35), 150). First values found: parameter(k3,10). parameter(k4,280). Parameter Search from LTL(R) Properties

22. François FagesShonan village 14/11/11

23. François FagesShonan village 14/11/11

24. François FagesShonan village 14/11/11

25. François FagesShonan village 14/11/11 LTL(R) Satisfaction Degree Bifurcation diagram on k4, k6 Continuous satisfaction degree in [0,1] [Tyson 91] of the LTL(R) formula for oscillation with amplitude constraint [Rizk Batt Fages Soliman CMSB 08]

26. François FagesShonan village 14/11/11

27. François FagesShonan village 14/11/11

28. François FagesShonan village 14/11/11

29. François FagesShonan village 14/11/11

30. François FagesShonan village 14/11/11

31. François FagesShonan village 14/11/11

32. François FagesShonan village 14/11/11

33. François FagesShonan village 14/11/11

34. François FagesShonan village 14/11/11

35. François FagesShonan village 14/11/11

36. François FagesShonan village 14/11/11

37. François FagesShonan village 14/11/11 [Rizk Batt Fages Soliman ISMB’09 Bioinformatics]

38. François FagesShonan village 14/11/11 Conclusion New focus in Systems Biology: formal methods from Computer Science –Beyond diagrammatic notations: formal semantics, abstract interpretation –Beyond simulation: symbolic execution –Beyond curve fitting: high-level specifications in temporal logic –Automatic model-checking. Parameter optimization. Model reduction New focus in Programming: numerical methods –Beyond discrete machines: stochastic or continuous or hybrid dynamics –Quantitative transition systems, hybrid systems –Continuous satisfaction degree of Temporal Logic specifications, optimization –From model checking (back) to model synthesis [Pnueli 77]

39. François FagesShonan village 14/11/11 Acknowledgments Contraintes group at INRIA Paris-Rocquencourt on this topic: Grégory Batt, Sylvain Soliman, Elisabetta De Maria, Aurélien Rizk, Steven Gay, Dragana Jovanovska, Faten Nabli, Xavier Duportet, Janis Ulhendorf EraNet SysBio C5Sys (follow up of FP6 Tempo) on cancer chronotherapies, coord. Francis Lévi, INSERM; Jean Clairambault INRIA; … Coupled models of cell and circadian cycles, p53/mdm2, cytotoxic drugs. INRIA/INRA project Regate coord. F. Clément INRIA; E. Reiter, D. Heitzler Modeling of GPCR Angiotensine and FSH signaling networks ANR project Calamar, coord. C. Chaouiya, D. Thieffry, ENS, L. Calzone, Curie... Modularity and Compositionality in regulatory networks. OSEO Biointelligence, coord. Dassault-Systèmes ANR Iceberg, coord Grégory Batt, with P. Hersen, Physics, O Gandrillon CNRS,… Model-based control of gene expression in microfluidic ANR Syne2arti, coord. Grégory Batt, INRIA with D. Drasdo INRIA, O. Maler, CNRS Verimag, R. Weiss MIT Model-based control of tissue growth in synthetic biology